- Email: [email protected]
An Imprinted Gene Underlies Postzygotic Reproductive Isolation in Arabidopsis thaliana
Developmental Cell
Article An Imprinted Gene Underlies Postzygotic Reproductive Isolation in Arabidopsis thaliana David Kradolfer,1,2 Philip Wolff,1,2 Hua Jiang,1 Alexey Siretskiy,1 and Claudia Ko¨hler1,* 1Department of Plant Biology and Forest Genetics, Uppsala BioCenter, Swedish University of Agricultural Sciences and Linnean Center of Plant Biology, 750 07 Uppsala, Sweden 2Department of Biology and Zu ¨ rich-Basel Plant Science Center, Swiss Federal Institute of Technology, ETH Centre, CH-8092 Zu¨rich, Switzerland *Correspondence: [email protected] http://dx.doi.org/10.1016/j.devcel.2013.08.006
SUMMARY
Postzygotic reproductive isolation in response to interploidy hybridizations is a well-known phenomenon in plants that forms a major path for sympatric speciation. A main determinant for the failure of interploidy hybridizations is the endosperm, a nutritious tissue supporting embryo growth, similar to the functional role of the placenta in mammals. Although it has been suggested that deregulated imprinted genes underpin dosage sensitivity of the endosperm, the molecular basis for this phenomenon remained unknown. In a genetic screen for suppressors of triploid seed abortion, we have identified the paternally expressed imprinted gene ADMETOS (ADM). Here, we present evidence that increased dosage of ADM causes triploid seed arrest. A large body of theoretical work predicted that deregulated imprinted genes establish the barrier to interploidy hybridization. Our study thus provides evidence strongly supporting this hypothesis and generates the molecular basis for our understanding of postzygotic hybridization barriers in plants.
INTRODUCTION How new species are formed remains a major question in evolutionary biology. Solving this question requires an understanding of the genetic and evolutionary basis of the traits mediating reproductive isolation. Postzygotic reproductive isolation in response to interploidy hybridization forms a major path for sympatric speciation in plants (Otto and Whitton, 2000; Ramsey and Schemske, 1998). The phenomenon has been termed ‘‘triploid block,’’ describing the difficulty of obtaining viable triploid seeds by diploid-tetraploid crosses (Marks, 1966). The endosperm is a key determinant for the success of interploidy hybridizations (Brink and Cooper, 1947; Woodell and Valentine, 1961), and abnormalities in the growth and structure of the endosperm are the source of the triploid block (Ramsey and Schemske, 1998). In most angiosperms, the endosperm is a triploid tissue derived after fertilization of the homodiploid central cell with a haploid sperm cell. The endosperm serves to support
and nurture the growing embryo, similar to the role of the placenta in mammals (Li and Berger, 2012). The endosperm of Arabidopsis thaliana follows the nuclear type of development where an initial syncytial phase of free nuclear divisions without cytokinesis is followed by cellularization (Costa et al., 2004). Endosperm cellularization is an important developmental transition required for viable seed formation (Hehenberger et al., 2012). Reciprocal hybridizations of plants that differ in ploidy have antagonistic effects on endosperm development; pollinations of maternal plants with pollen donors of higher ploidy will cause delay or failure of endosperm cellularization correlating with increased seed size or seed abortion. Conversely, the reciprocal cross will cause precocious endosperm cellularization and decreased seed size (Scott et al., 1998). In Arabidopsis, hybridizations of a diploid female plant with a tetraploid pollen donor (paternal excess hybridizations) will cause the formation of triploid progeny that will abort in an accession-dependent frequency. Although abortion of triploid seeds is highly penetrant in the Arabidopsis Columbia (Col) accession, the response to interploidy hybridization is rather weak in the Landsberg erecta (Ler) and C24 accessions (Scott et al., 1998; Dilkes et al., 2008). The Endosperm Balance Number (EBN) serves as a predictor for the success of interploidy hybridizations by assigning each species a defined effective ploidy that must be in a 2:1 maternal to paternal ratio in the endosperm for crosses to succeed (Johnston et al., 1980). Although widely used in plant breeding, the molecular mechanisms underpinning the EBN concept remained enigmatic. The sensitivity of the endosperm to changes in parental genome balance led to the hypothesis that imprinted genes are causal for the response to interploidy hybridizations (Haig and Westoby, 1989; Moore and Haig, 1991; Birchler, 1993; Gutierrez-Marcos et al., 2003; Zeh and Zeh, 2000; Kinoshita, 2007). Imprinted genes are predominantly expressed from either maternally or paternally inherited alleles as a result of epigenetic modifications established in the gametes (Ko¨hler et al., 2012). If imprinted genes encode for dosage-sensitive regulators, disturbed balance of these regulators in response to interploidy crosses is expected to cause deleterious effects (Birchler et al., 2001; Veitia, 2003; Birchler and Veitia, 2012). Although previous studies revealed deregulation of imprinted genes in response to interploidy hybridizations that were either connected with the activation of the silenced allele or hyperactivation of the active allele (Jullien and Berger, 2010; Tiwari et al., 2010; Wolff et al.,
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 525
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
2011), whether indeed there is a causal relationship between deregulated imprinted genes and seed abortion remained unknown. In this study, we have identified the paternally expressed imprinted gene (PEG) ADMETOS (ADM) to be causally responsible for abortion of triploid paternal excess Arabidopsis seeds. ADM acts as a dosage-sensitive regulator of triploid seed development and increased ADM expression in triploid seeds causes endosperm cellularization failure leading to seed abortion. Our study establishes ADM as a dosage-sensitive imprinted gene that erects a strong postzygotic reproductive barrier upon change of genome dosage. RESULTS Identification of a Suppressor of Triploid Seed Abortion To identify the genetic basis of interploidy hybridization barriers we searched for mutants that act as suppressors of triploid seed abortion. We made use of the jason (jas) mutant that frequently forms unreduced (2n) pollen and triploid seeds upon self-fertilization (Erilova et al., 2009). We generated a mutagenized population of diploid jas-3 plants (in the Col background) and screened for mutants with a reduced seed abortion rate. We identified one mutant that strongly suppressed triploid seed abortion and that we named admetos (adm-1) after one of Jason’s companions in his quest for the Golden Fleece. Although jas-3 formed about 30% aborted triploid seeds, only about 2% of seeds aborted in the jas-3 adm-1 double mutant (Figures 1A and 1C). We tested viability of jas-3 adm-1 seeds and found that 97% of seeds germinated compared with only 80% of jas-3 seeds (Figure 1B). Rescued triploid adm-1 seeds were enlarged compared to diploid wildtype seeds (Figure 1C). We tested ploidy of enlarged seeds and found 11 out of 12 tested seeds to be triploid (data not shown). We tested whether the ability of adm-1 to rescue triploid seeds is independent of the mutation giving rise to unreduced gametes. For that purpose, we introduced adm-1 into the omission of second division 1 (osd1) mutant background. In contrast to jas, which produces gametes by a first division restitution mechanism, the osd1 mutant forms unreduced gametes by a second division restitution mechanism (d’Erfurth et al., 2009). The osd1 mutation is highly penetrant and leads to almost 100% 2n pollen and triploid seeds when used as pollen donor (d’Erfurth et al., 2009). Similar to the rescue of triploid seeds in the jas-3 background, adm-1 rescued triploid seeds formed by unreduced pollen of the osd1 mutant (Figures 1D–1F). The majority of triploid adm-1 seeds formed by unreduced pollen of the osd1 mutant germinated and developed into seedlings (Figures 1E and 1G). In contrast to triploid seeds that fail to undergo endosperm cellularization (Scott et al., 1998), the endosperm of triploid adm-1 seeds cellularized, albeit delayed compared to diploid wild-type seeds (Figure 1H). Whereas wild-type seeds cellularize 5–7 days after pollination (DAP), triploid adm-1 seeds cellularized 8–10 DAP. Endosperm cellularization failure is likely responsible for embryo arrest (Hehenberger et al., 2012). In agreement with that view, restoration of endosperm cellularization in triploid adm-1 seeds allows progression of embryo development to the mature embryo stage (Figure 1H).
Expression of AGL MADS-Box Genes and Imprinted Genes Is Normalized in Triploid adm Seeds Previous studies revealed that failure of endosperm cellularization was associated with increased expression of AGAMOUS LIKE (AGL) MADS-box transcription factor genes (Erilova et al., 2009; Walia et al., 2009; Jullien and Berger, 2010; Hehenberger et al., 2012). Consistently, normalization of endosperm cellularization and embryo development by adm-1 was reflected by normalization of AGL gene expression (Figure 2A). Expression of AGL36, AGL62, and AGL90 was close to wild-type expression levels in triploid adm-1 seeds, whereas PHE1 and AGL40 had reduced expression levels in triploid adm-1 seeds compared to wild-type triploid seeds but still significantly increased expression compared to diploid wild-type seeds. Expression of AGL28 was not affected by the adm-1 mutation (Figure 2A). We also tested expression of MEA and FIS2 that are inversely deregulated in Col triploid seeds (Jullien and Berger, 2010). Whereas expression of MEA was reduced to wild-type levels in adm-1 triploid seeds, FIS2 expression was increased in adm-1 triploid seeds compared to diploid and triploid wild-type seeds (Figure 2B). Unbalanced parental contributions in the endosperm cause deregulated expression of imprinted genes (Jullien and Berger, 2010; Wolff et al., 2011). We therefore tested whether adm-1 could normalize expression of maternally and paternally expressed imprinted genes (MEGs and PEGs, respectively) that have increased expression levels in response to interploidy crosses (Wolff et al., 2011). Indeed, all tested PEGs had reduced expression levels in adm-1 triploid seeds compared to triploid seeds expressing a wild-type ADM allele (Figure 2C). Three out of six tested MEGs had increased expression levels in triploid seeds and their expression was decreased in adm-1 triploid seeds (Figure 2D). ADM Is a Paternally Expressed Imprinted Gene To examine whether adm-1 has a maternal or paternal effect on triploid seed rescue, we performed reciprocal crosses of adm-1 in the osd1 mutant background. These crosses revealed that adm-1 confers seed rescue when paternally but not when maternally inherited (Figures 3A and 3B). Loss of ADM affected only development of triploid but not diploid seeds. Seed development, final seed size, and expression of selected AGL genes were similar in diploid mutant seeds and wildtype seeds, suggesting a specific negative effect of increased ADM dosage in triploid seeds (Figures S1A–S1G available online). We identified the adm-1 mutation to cause an amino acid substitution in gene At4g11940, encoding for a protein belonging to the diverse family of molecular chaperones called J-domain proteins (Rajan and D’Silva, 2009; Kampinga and Craig, 2010). The mutation is located within the J-domain and changes an arginine at position 81 to a tryptophan, which likely impairs the function of the protein (Figures S2A and S2B). Consistently, genetic analysis revealed that adm-1 is likely a recessive mutant because diploid jas-3 pollen heterozygous for adm-1 did not produce viable triploid seeds (Figure S2C). Although all J-domain proteins share the name-giving conserved J-domain, they differ by the presence of different carboxyterminal domains. ADM belongs to the type IV class of J-domain proteins, which contain a J-domain lacking the
526 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 1. Triploid adm Seeds Are Viable (A and D) Percentage of aborted seeds of self-fertilized jas-3 and jas-3 adm-1 mutants (A) and Col wild-type plants pollinated with osd1 (D) and adm-1 mutants pollinated with osd1 adm-1 pollen (D). Numbers on top of bars correspond to number of analyzed seeds. (B and E) Percentage of germinated seeds of indicated genotypes. Numbers on top of bars correspond to number of analyzed seeds. (C and F) Pictures of mature seeds of indicated genotypes. Selected triploid seeds are marked with arrowheads. Scale bar, 1 mm. (G) Seedlings 6 days after germination. Scale bar, 1 cm. (H) Sections of seeds with the indicated genotype at 8 DAP (upper panels) and 10 DAP (lower panels). Ploidy was inferred based on seed morphology. Scale bar, 0.1 mm. See also Figures S1 and S2.
characteristic HPD motif required for stimulation of HSP70’s ATPase activity (Rajan and D’Silva, 2009; Kampinga and Craig, 2010). J-domain proteins are involved in multiple biological processes such as protein folding, protein translocation into
organelles, and regulation of gene expression by assisting chromatin regulators (Dai et al., 2005; Kampinga and Craig, 2010; Wang and Brock, 2003). Orthologous sequences to ADM were only identified in species of the Brassicaceae
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 527
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 2. Expression of AGL MADS-Box Genes and Imprinted Genes Is Decreased in adm Quantitative RT-PCR analysis of indicated genes in whole siliques derived from crosses of wild-type Col plants (white bars), wild-type Col plants pollinated with osd1 pollen (gray bars), and adm-1 plants pollinated with osd1 adm-1 pollen (black bars) at 6 DAP. (A) AGAMOUS LIKE (AGL) genes. (B) MEA and FIS2. (C) Paternally expressed imprinted genes. (D) Maternally expressed imprinted genes. Error bars indicate SEM.
closely related to Arabidopsis (A. lyrata, A. arenosa, and Capsella rubella; Figure S2A) but not in more distantly related species such as Brassica rapa nor in species outside the Brassicaceae. At4g11940 has previously been predicted to be a paternally expressed imprinted gene (Hsieh et al., 2011; Wolff et al., 2011), which is in agreement with our genetic data that reveal a functional role of ADM when paternally inherited. We introduced a construct containing the promoter and coding region of At4g1190 into the osd1 adm-1 background and tested whether it could complement the adm-1 phenotype. Indeed, the fraction of aborted seeds was increased to about 85% when we pollinated adm-1 plants with osd1
adm-1 pollen homozygous for the At4g11940 construct (Figure S2D). ADM was expressed during seed development, reaching highest expression levels at 3 DAP (Figure 3C). We generated a translational reporter of ADM with GFP and tested expression in pollen and developing seeds (Figures 3D–3K). The ADM::ADM-GFP transgene was functional as revealed by increased rates of seed abortion of triploid adm-1 seeds (Figure S2D). ADM was specifically localized in endosperm nuclei (Figures 3D–3K). Reciprocal crosses revealed that the ADM-GFP transgene was only expressed from the paternal allele (Figures 3H and 3I). We did not detect expression of ADM in pollen (Figures 3J and 3K) and neither in other tissues (data not shown), revealing that expression of the paternal ADM allele is specifically activated in the endosperm. ADM transcript levels were strongly increased in triploid seeds (Figure 3L). We tested whether increased ADM transcript levels are caused by breakdown of ADM imprinting and activation of the maternal ADM allele. However, ADM remained imprinted in triploid seeds (Figure 3M), revealing a specific effect of double paternal genome dosage on expression of the paternal ADM allele. ADM Is a Dosage-Sensitive Gene We addressed the question whether ADM is a dosage-sensitive gene, which would imply that triploid seed rescue correlates with decreasing ADM gene dosage. In agreement with this hypothesis, there was a decreased proportion of aborted triploid seeds when adm-1 was biparentally inherited compared to paternal inheritance (Figure 3B), suggesting a functional role of the maternal ADM alleles in triploid seeds. In agreement with that, there were residual levels of the maternal ADM alleles detectable in diploid and triploid seeds (Figure 3M;
528 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 3. ADM Is a Paternally Expressed Imprinted Gene and Remains Imprinted in Triploid Seeds (A) Pictures of mature seeds of adm-1 plants pollinated with osd1 pollen and wild-type Col plants pollinated with osd1 adm-1 pollen. (B) Percentages of aborted and germinated seeds of wild-type Col plants pollinated with osd1 pollen (white bars), adm-1 plants pollinated with osd1 pollen (light gray bars), wild-type Col plants pollinated with osd1 adm-1 pollen (dark gray bars), and adm-1 plants pollinated with osd1 adm-1 pollen (black bars). Numbers on top of bars correspond to number of analyzed seeds. (C) Expression of ADM in whole siliques from 1–6 DAP. Error bars indicate SEM. (D–G) ADM::ADM-GFP expression in self-fertilized seeds at 1, 2, 3, and 4 DAP. (H and I) Reciprocal crosses with ADM::ADM-GFP transmitted through pollen (H) and female gametes (I) at 4 DAP. (J and K) Only background fluorescence is detectable in ADM::ADM-GFP pollen (J) and wild-type pollen (K). (L) Expression of ADM in whole siliques derived from crosses of wild-type Col plants and wild-type plants pollinated with osd1 pollen at 6 DAP. Error bars indicate SEM. (M) ADM is predominantly paternally expressed in diploid and triploid seeds. Siliques of crosses of Ler plants pollinated with Col wild-type or jas-3 pollen were harvested at the indicated time points and allele-specific expression was tested by restriction-based allele-specific PCR analysis. Siliques of Col and Ler plants pollinated with Col and Ler pollen were used as controls. Scale bars, 1 mm (A) and 50 mm (D–K). See also Figure S3.
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 529
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 4. ADM Is a Dosage-Sensitive Gene (A) Distribution of seed sizes from self-fertilized wild-type Col plants (blue line), wild-type Col plants pollinated with osd1 pollen (red line), adm-1 plants pollinated with osd1 pollen (green line), wild-type Col plants pollinated with osd1 adm-1 pollen (purple line), and adm-1 plants pollinated with osd1 adm-1 pollen (black line). A minimum of 100 seeds was analyzed for each cross. (B) Frequencies of aborted and germinated triploid seeds derived from wild-type Col plants pollinated with osd1 pollen (light gray bars), wildtype Col plants pollinated with osd1 adm-2 pollen (dark gray bars), and adm-2 plants pollinated with osd1 adm-2 pollen (black bars). Numbers above the bars indicate the number of analyzed seeds. (C) Pictures of seeds of the indicated crosses. Scale bar, 1 mm. (D) Expression of AGL genes in whole siliques of indicated genotypes at 6 DAP. Error bars indicate SEM. See also Figure S4.
Figure S3). A functional role of maternal ADM expression, however, became unmasked only upon loss of paternal ADM alleles. Triploid seeds that were homozygous for adm-1 were decreased in size compared to triploid seeds inheriting a wild-type maternal ADM allele (genotype adm-1/ /+), revealing a dosage-sensitive role of ADM in regulating seed size (Figure 4A). We identified a second adm allele (adm-2; Figures S4A–S4C) in a collection of T-DNA mutants that similarly to adm-1 strongly suppressed triploid seed abortion (Figures 4B and 4C). In agreement with ADM being a dosage-sensitive gene, also adm-2 had a stronger effect on triploid seed viability when the seeds were homozygous for the adm-2 mutation compared to triploid seeds having a functional ADM maternal
allele (Figures 4B and 4C). We tested whether ADM dosage-dependent abortion of triploid seeds was reflected on transcript levels of AGL genes. Indeed, transcript levels of AGL genes in triploid adm-2 seeds were partially normalized (Figure 4D) with the clear trend that AGL expression was more strongly reduced in homozygous adm-2 triploid seeds than in triploid seeds that inherited an ADM maternal allele. The repressive effect of adm-2 on AGL gene expression was detectable at 4 DAP (Figure S4D), clearly preceding the onset of endosperm cellularization at 6 DAP (Kradolfer et al., 2013). Therefore, reduced AGL gene expression by adm in triploid seeds is unlikely a secondary consequence of normalized endosperm cellularization, but rather enables endosperm cellularization. Taken together, these data reveal that ADM is a dosage-sensitive gene and show that the maternal ADM allele gains a functional role in triploid seeds. We addressed the question whether ADM overexpression would be sufficient to confer diploid seed abortion. We generated transgenic lines overexpressing ADM under the control of the endosperm-specific PHE1 promoter (Weinhofer et al., 2010) and the RPS5A promoter, which is expressed in embryo and endosperm (Weijers et al., 2001). We generated 10 independent transgenic lines for each construct and identified several lines with medium to strong overexpression in the endosperm (Figure S4E). However, in none of these lines did we detect triploid seed-like abortion, revealing that increased ADM expression is specifically detrimental in triploid seeds.
530 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
in triploid Ler seeds compared to triploid Col seeds (Figure 5A), correlating with a reduced abortion rate of triploid Ler seeds. Introducing a Ler genomic locus of ADM (including 1.5 kb of upstream and 0.3 kb of downstream sequence) into the adm-1 mutant background (Col accession) complemented the mutant similarly well as an ADM genomic fragment derived from the Col accession (Figure 5B), revealing that ADM sequence variation between Col and Ler accessions (Figure S5) is unlikely to be responsible for accession-dependent differences in triploid seed abortion. In support of this conclusion, expression of a Col ADM genomic fragment in the jas-1 mutant (Ler accession) did not increase the frequency of triploid seed abortion (Figure 5C). Together, these data strongly suggest that the different triploid responses of Col and Ler are unlikely consequences of ADM sequence variation in cis but rather are caused by accession-specific factors acting in trans on ADM expression.
Figure 5. Accession-Dependent Triploid Seed Abortion Correlates with ADM Expression Levels (A) Expression of ADM in whole siliques derived from pollinations of wild-type Col or Ler plants with wild-type Col or Ler pollen (white and dark gray bars, respectively), Col plants pollinated with jas-3 pollen (Col background; light gray bars), and Ler plants pollinated with jas-1 pollen (Ler background; black bars) at 4 and 6 DAP. Error bars indicate SEM. (B) adm-1 plants were pollinated with pollen of osd1, osd1 adm-1, or osd1 adm-1 mutants hemizygous for a Col or Ler ADM::ADM transgene. Percentage of collapsed seeds is shown. Numbers above the bars indicate the number of analyzed seeds. Independent transgenic lines are numbered (#). (C) A Col ADM::ADM transgene was expressed in the jas-1 (Ler) background. Percentage of collapsed seeds for self-fertilized plants is shown. Numbers above the bars indicate the number of analyzed seeds. Independent transgenic lines are numbered (#). See also Figure S5.
Accession-Dependent Seed Abortion Correlates with ADM Expression Levels Triploid seed abortion is accession dependent. Although triploid seeds abort very frequently in the Col accession, the majority of triploid seeds are viable in the Ler background (Dilkes et al., 2008). We addressed the question whether accession-dependent differences of Col and Ler are caused by different expression or function of ADM. Therefore, we tested whether triploid Col and Ler seeds differ in ADM expression levels. Indeed, we found that ADM transcript levels remained at a much lower level
adm Partially Rescues Seeds Lacking FIS-PRC2 Loss of function of the FERTILIZATION INDEPENDENT SEED (FIS) Polycomb Repressive Complex 2 (PRC2) causes a similar phenotype as paternal excess triploid seeds, correlating with largely overlapping sets of deregulated genes (Erilova et al., 2009; Tiwari et al., 2010). Consistently, ADM is a target gene of the FIS-PRC2 in the endosperm (Figure S6), and ADM expression was increased in mutants of the FIS-PRC2 subunit MEA (Figure 6A). We tested whether similar to its ability to rescue triploid seeds adm-1 could rescue mea mutant seeds. We generated double mutants that were heterozygous for mea and homozygous for adm-1 and pollinated them with adm-1 mutant pollen. Whereas a mea/+ ADM/+ mutant formed about 50% aborting seeds (n = 216), seed abortion in mea/+ adm-1/ double mutants was suppressed to 36% (n = 357; X2 = 29.14 > X0.05[1]2 = 3.84). Among seeds produced by a mea/+ adm-1/ mutant, we found 10% partially collapsed seeds that contained a developed embryo (Figures 6B and 6C). We tested viability of aborted and partially collapsed seeds by germinating them on plates and genotyping the seedlings for the presence of the mea mutation. We found that 15% of mea/+ adm-1/ seeds germinated, compared to only 1%–3% of mea/+ and mea/+ adm-1/+ seeds (Figure 6D). These data reveal that loss of ADM function allows the requirement of FIS-PRC2 function to be bypassed. DISCUSSION Here, we report the identification of the imprinted ADM gene, which determines seed viability in paternal excess interploidy hybridizations. Whereas triploid seeds abort at a frequency close to 100% in the Col accession, mutation of ADM leads to nearly complete seed rescue. Triploid adm seeds have reduced expression levels of AGL genes, including AGL62, a central regulator of endosperm cellularization. Consistently, cellularization is restored in triploid adm seeds. The parental conflict theory predicts that genomic imprinting evolved as a consequence of intragenomic conflict over nutrient allocation from the mother to the offspring (Haig and Westoby, 1989). Based on this prediction, PEGs should have a predominant role in supporting growth of tissues that sustain embryo growth (placenta and endosperm) (Haig and Westoby, 1989;
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 531
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 6. adm Partially Suppresses Abortion of mea Mutant Seeds (A) Quantitative RT-PCR analysis of ADM in whole siliques derived from pollinations of wild-type Col plants or mea/+ mutants with wild-type Col pollen at 6 DAP. Error bars indicate SEM. (B) Percentages of aborted, partially collapsed, and normal seeds derived from pollinations of mea/+ with wild-type Col pollen, mea/+ adm-1 with wild-type Col pollen, mea/+ with adm-1 pollen, and mea/+ adm-1 with adm-1 pollen (left). Numbers above the bars indicate the number of analyzed seeds. Examples of seeds of the three categories (right panels). Scale bar, 0.5 mm. (C) Clearings of seeds derived from pollinations of mea/+ adm-1 with adm-1 pollen at 12 DAP. Wild-type seed (left), aborting mea seed (middle), and rescued mea seed (right) are segregated. Scale bar, 0.1 mm. (D) Transmission frequency of the mea mutant allele through female gametes was determined by genotyping germinated seedlings for the presence of the mea mutation. See also Figure S6.
Smith et al., 2006). Our results provide substantial support to this hypothesis by revealing that the PEG ADM has a role in promoting endosperm growth by regulating endosperm cellularization in triploid seeds. The effect on endosperm cellularization is likely a direct or indirect consequence of ADM, regulating expression of AGL genes in triploid seeds. Previous studies revealed that reduced paternal genome contribution can bypass the requirement of the FIS-PRC2 for viable seed formation (Nowack et al., 2007; Kradolfer et al., 2013). Here, we show that reduced dosage of ADM can substantially contribute to fis mutant rescue, thus providing strong support for the view that the FIS-PRC2 regulates viable seed formation by regulating expression of PEGs. Whereas increased expression of ADM is causally responsible for triploid seed abortion, increased expression of ADM did not cause major changes in diploid seed development. These data reveal that changes in ADM dosage have different consequences in diploid and triploid seeds. Moreover, impaired ADM function did not cause noticeable aberrations in diploid seed development, suggesting that there is either residual ADM function in adm-1 and adm-2 mutants, or, alternatively, that ADM is not essential for seed development. The last scenario bears
striking similarities to the lethal effect mediated by Lethal hybrid rescue (Lhr) in F1 hybrid sons of Drosophila melanogaster and D. simulans (Brideau et al., 2006). A loss-of-function mutation at the D. simulans Lhr locus can reverse F1 lethality; however, neither does increased expression of the D. simulans Lhr cause lethality in D. melanogaster nor does loss of Lhr function cause a negative effect (Chatterjee et al., 2007; Brideau and Barbash, 2011). Therefore, the Lhr protein acquires a negative ectopic interaction in the hybrid background, similar to the negative effect of ADM in triploid seeds. Speciation is characterized by the evolution of barriers to reproductive exchange between two groups of organisms. The genes building these hybridization barriers are speciation genes. Our study identifies ADM as an imprinted plant speciation gene that underlies the hybridization barrier between diploid and tetraploid Arabidopsis plants. Although only a few natural Arabidopsis thaliana tetraploid accessions have been identified (Bomblies et al., 2007), unreduced male gametes are frequently formed after cold stress (De Storme et al., 2012). It is generally assumed that the formation of stable tetraploids occurs via the formation of unreduced gametes and unstable triploid intermediates (Rieseberg and Willis, 2007). We
532 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
therefore propose that suppression of triploid seed formation by ADM contributes to the low frequency of natural Arabidopsis thaliana tetraploid populations. However, because the triploid block mediated by ADM is active only in response to increased paternal genome contribution, triploid seed formation can occur when unreduced gametes are maternally formed, explaining the existence of tetraploid populations of Arabidopsis thaliana and relatives. A large body of theoretical work led to the prediction that imprinted genes erect hybridization barriers (Haig and Westoby, 1989; Moore and Haig, 1991; Birchler, 1993; Zeh and Zeh, 2000; Gutierrez-Marcos et al., 2003; Kinoshita, 2007). This theory builds on the hypothesis that imprinted genes are dosage sensitive and imprinted expression is disrupted in response to hybridization. Disruption of imprinted gene expression has been shown to occur at some selected loci in response to interploidy and interspecies hybridization in Arabidopsis (Josefsson et al., 2006; Jullien and Berger, 2010) as well as in the North American rodent species Peromyscus (Vrana et al., 2000). Our study shows that interploidy hybridizations result in hyperactivation of the active ADM allele while not affecting the imprinted status of ADM. Until now experimental evidence that unbalanced expression of an imprinted gene can form a hybridization barrier was missing. Thus, the identification of ADM as a dosage-sensitive imprinted gene represents a major advance because it is an imprinted plant gene that erects a strong postzygotic reproductive barrier upon change of genome dosage. One implication of our data is that alterations in epigenetic gene regulation contribute to both the establishment and maintenance of reproductive isolation barriers in plants. EXPERIMENTAL PROCEDURES Plant Material and Growth Conditions Plants were grown in a growth chamber at 60% humidity and daily cycles of 16 hr light at 21 C and 8 hr darkness at 18 C. Arabidopsis thaliana mutants jas-3 (Erilova et al., 2009), mea-8 (Ngo et al., 2007), and adm-2 (SAIL_1252_D03) are in the Col accession. The osd1-1 mutant (d’Erfurth et al., 2009) was kindly provided by Raphael Mercier. The mutant was originally identified in the Nossen background and subsequently introgressed into Col by repeated backcrossing over five generations. The jas-1 mutant is in the Ler background (Erilova et al., 2009). For crosses, designated female partners were emasculated, and the pistils hand-pollinated 2 days after emasculation. RNA Extraction and qPCR Analysis Three siliques were harvested for each cross and frozen together with glass beads (1.25–1.55 mm; Carl Roth) in liquid nitrogen. The samples were ground in a Silamat S5 (IvoclarVivadent) two times for 6 s. RNA was extracted using the RNeasy Plant Mini Kit (QIAGEN) according to the manufacturer’s instructions. Residual DNA was removed using the QIAGEN RNase-free DNase Set, and cDNA was synthesized using the Fermentas First strand cDNA synthesis kit according to the manufacturer’s instructions. Quantitative real-time PCR was performed using an iQ5 Real-Time PCR Detection System (BioRad) and Maxima SYBR green qPCR master mix (Fermentas) according to the manufacturer’s instructions. Quantitative realtime PCR was performed with three replicates using primers as indicated in the Supplemental Experimental Procedures, and results were analyzed as described (Simon, 2003) using PP2A as a reference gene. EMS Mutagenesis and Mapping Diploid jas-3 seeds were selected based on seed size and mutagenized for 15 hr in 0.3% ethyl methanesulfonate (EMS). The seeds were germinated
on soil, and one side branch of each plant was harvested. We screened 900 M2 families consisting each of eight plants derived from one M1 plant and selected mutants with reduced seed abortion. For genetic mapping of the adm mutation, we established an F2 mapping population by crossing jas-3 adm with jas-3. From 896 analyzed plants, 200 had a clear adm phenotype and were selected for mapping. Leaf samples of all plants were pooled and DNA was extracted using the Nucleon PhytoPure Kit (Amersham Biosciences). Sequencing was performed on one lane of an Illumina Hiseq 2000 device using paired end mode at 70 fold coverage. Read alignment was done using the Bowtie algorithm (Langmead et al., 2009), and Samtools was employed for SNP calling (Li et al., 2009). Mapping using SNP data files was performed as described by (Austin et al., 2011). Microscopy For tissue sections, seeds were fixed and embedded with Technovit 7100 (Heraeus Kulzer) as described (Hehenberger et al., 2012). Five-micrometer sections were prepared with an HM 355 S microtome (Microm) using glass knives. Sections were stained for 1 min with 0.1% toluidine blue and washed three times with distilled water. For clearings, siliques were slit along the lamella and fixed overnight at 4 C in ethanol/acetic acid in a ratio of 9:1 and washed with 70% ethanol. Seeds were cleared for 1 day at 4 C in chloral hydrate solution (glycerol/chloral hydrate/water in a ratio of 1:8:3). Seeds and pollen of the GFP lines were mounted in water and analyzed with epifluorescence optics. Microscopy was performed using a Leica DMI 4000B microscope with DIC optics. Images were captured using a Leica DFC360 FX camera (Leica) and processed using Adobe Photoshop CS5. Pictures of mature seeds were taken using a Leica Z16apoA microscope and a Leica DFC425C camera and processed using Adobe Photoshop CS5. Flow Cytometry Ploidy levels were measured as described (Kradolfer et al., 2013). Seed Size Analysis Seeds were arranged on glass slides and pictures were taken using a Leica Z16apoA microscope with transmitted light. Seed size was measured as described by (Herridge et al., 2011). Germination and Transmission Analysis Seeds were surface sterilized using chlorine gas. Five milliliters of hydrochloric acid was added to 30 ml of sodium hypochlorite under a fume hood. This mixture was incubated in a container together with the seeds for 1 hr. Seeds were plated on MS media containing 1% sugar. After stratification at 4 C in the dark for 2 days, seedlings were grown in a growth room under a longday photoperiod (16 hr light and 8 hr darkness) at 23 C. Germination frequency was determined after 7 days. For transmission analysis, seedlings were harvested after 12 days for genotyping using primers specified in the Supplemental Experimental Procedures. Generation of Plasmids and Transgenic Lines Details on the generation of plasmids and transgenic lines are provided in the Supplemental Experimental Procedures. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures and six figures and can be found with this article online at http://dx.doi.org/ 10.1016/j.devcel.2013.08.006. ACKNOWLEDGMENTS We thank Sabrina Huber and Cecilia Wa¨rdig for excellent technical support, Sofia Berlin Kolm and Juliette de Meaux for helpful discussions, and Lars Hennig and members of the Ko¨hler laboratory for critical comments on this manuscript. This research was supported by a fellowship of the Zu¨rich-Basel Plant Science Center (to D.K.) and a European Research Council Starting Independent Researcher grant (to C.K.).
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 533
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Received: April 3, 2013 Revised: June 18, 2013 Accepted: August 10, 2013 Published: September 5, 2013 REFERENCES Austin, R.S., Vidaurre, D., Stamatiou, G., Breit, R., Provart, N.J., Bonetta, D., Zhang, J., Fung, P., Gong, Y., Wang, P.W., et al. (2011). Next-generation mapping of Arabidopsis genes. Plant J. 67, 715–725. Birchler, J.A. (1993). Dosage analysis of maize endosperm development. Annu. Rev. Genet. 27, 181–204. Birchler, J.A., Bhadra, U., Bhadra, M.P., and Auger, D.L. (2001). Dosagedependent gene regulation in multicellular eukaryotes: implications for dosage compensation, aneuploid syndromes, and quantitative traits. Dev. Biol. 234, 275–288. Birchler, J.A., and Veitia, R.A. (2012). Gene balance hypothesis: connecting issues of dosage sensitivity across biological disciplines. Proc. Natl. Acad. Sci. USA 109, 14746–14753. Bomblies, K., Lempe, J., Epple, P., Warthmann, N., Lanz, C., Dangl, J.L., and Weigel, D. (2007). Autoimmune response as a mechanism for a DobzhanskyMuller-type incompatibility syndrome in plants. PLoS Biol. 5, e236. Brideau, N.J., and Barbash, D.A. (2011). Functional conservation of the Drosophila hybrid incompatibility gene Lhr. BMC Evol. Biol. 11, 57. Brideau, N.J., Flores, H.A., Wang, J., Maheshwari, S., Wang, X., and Barbash, D.A. (2006). Two Dobzhansky-Muller genes interact to cause hybrid lethality in Drosophila. Science 314, 1292–1295. Brink, R., and Cooper, D. (1947). The endosperm in seed development. Bot. Rev. 132, 479–541. Chatterjee, R.N., Chatterjee, P., Pal, A., and Pal-Bhadra, M. (2007). Drosophila simulans Lethal hybrid rescue mutation (Lhr) rescues inviable hybrids by restoring X chromosomal dosage compensation and causes fluctuating asymmetry of development. J. Genet. 86, 203–215. Costa, L.M., Gutie`rrez-Marcos, J.F., and Dickinson, H.G. (2004). More than a yolk: the short life and complex times of the plant endosperm. Trends Plant Sci. 9, 507–514. Dai, Y.S., Xu, J., and Molkentin, J.D. (2005). The DnaJ-related factor Mrj interacts with nuclear factor of activated T cells c3 and mediates transcriptional repression through class II histone deacetylase recruitment. Mol. Cell. Biol. 25, 9936–9948. De Storme, N., Copenhaver, G.P., and Geelen, D. (2012). Production of diploid male gametes in Arabidopsis by cold-induced destabilization of postmeiotic radial microtubule arrays. Plant Physiol. 160, 1808–1826. d’Erfurth, I., Jolivet, S., Froger, N., Catrice, O., Novatchkova, M., and Mercier, R. (2009). Turning meiosis into mitosis. PLoS Biol. 7, e1000124. Dilkes, B.P., Spielman, M., Weizbauer, R., Watson, B., Burkart-Waco, D., Scott, R.J., and Comai, L. (2008). The maternally expressed WRKY transcription factor TTG2 controls lethality in interploidy crosses of Arabidopsis. PLoS Biol. 6, 2707–2720. Erilova, A., Brownfield, L., Exner, V., Rosa, M., Twell, D., Mittelsten Scheid, O., Hennig, L., and Ko¨hler, C. (2009). Imprinting of the polycomb group gene MEDEA serves as a ploidy sensor in Arabidopsis. PLoS Genet. 5, e1000663. Gutierrez-Marcos, J.F., Pennington, P.D., Costa, L.M., and Dickinson, H.G. (2003). Imprinting in the endosperm: a possible role in preventing wide hybridization. Philos. Trans. R. Soc. Lond. B Biol. Sci. 358, 1105–1111. Haig, D., and Westoby, M. (1989). Parent-specific gene expression and the triploid endosperm. Am. Nat. 134, 147–155. Hehenberger, E., Kradolfer, D., and Ko¨hler, C. (2012). Endosperm cellularization defines an important developmental transition for embryo development. Development 139, 2031–2039.
development and expression of paternally inherited alleles. Development 134, 4107–4117. Herridge, R.P., Day, R.C., Baldwin, S., and Macknight, R.C. (2011). Rapid analysis of seed size in Arabidopsis for mutant and QTL discovery. Plant Methods 7, 3. Hsieh, T.F., Shin, J., Uzawa, R., Silva, P., Cohen, S., Bauer, M.J., Hashimoto, M., Kirkbride, R.C., Harada, J.J., Zilberman, D., and Fischer, R.L. (2011). Regulation of imprinted gene expression in Arabidopsis endosperm. Proc. Natl. Acad. Sci. USA 108, 1755–1762. Johnston, S., Nijs, T., Peloquin, S., and Hanneman, R., Jr. (1980). The significance of genic balance to endosperm development in interspecific crosses. Theor. Appl. Genet. 57, 5–9. Josefsson, C., Dilkes, B., and Comai, L. (2006). Parent-dependent loss of gene silencing during interspecies hybridization. Curr. Biol. 16, 1322–1328. Jullien, P.E., and Berger, F. (2010). Parental genome dosage imbalance deregulates imprinting in Arabidopsis. PLoS Genet. 6, e1000885. Kampinga, H.H., and Craig, E.A. (2010). The HSP70 chaperone machinery: J proteins as drivers of functional specificity. Nat. Rev. Mol. Cell Biol. 11, 579–592. Ko¨hler, C., Wolff, P., and Spillane, C. (2012). Epigenetic mechanisms underlying genomic imprinting in plants. Annu. Rev. Plant Biol. 63, 331–352. Kradolfer, D., Hennig, L., and Ko¨hler, C. (2013). Increased maternal genome dosage bypasses the requirement of the FIS polycomb repressive complex 2 in Arabidopsis seed development. PLoS Genet. 9, e1003163. Langmead, B., Trapnell, C., Pop, M., and Salzberg, S.L. (2009). Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25. Li, J., and Berger, F. (2012). Endosperm: food for humankind and fodder for scientific discoveries. New Phytol. 195, 290–305. Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., and Durbin, R.; 1000 Genome Project Data Processing Subgroup. (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079. Marks, G. (1966). The origin and significance of intraspecific polyploidy: experimental evidence from Solanum chacoense. Evolution 20, 552–557. Moore, T., and Haig, D. (1991). Genomic imprinting in mammalian development: a parental tug-of-war. Trends Genet. 7, 45–49. Nowack, M.K., Shirzadi, R., Dissmeyer, N., Dolf, A., Endl, E., Grini, P.E., and Schnittger, A. (2007). Bypassing genomic imprinting allows seed development. Nature 447, 312–315. Otto, S.P., and Whitton, J. (2000). Polyploid incidence and evolution. Annu. Rev. Genet. 34, 401–437. Rajan, V.B., and D’Silva, P. (2009). Arabidopsis thaliana J-class heat shock proteins: cellular stress sensors. Funct. Integr. Genomics 9, 433–446. Ramsey, J., and Schemske, D. (1998). Pathways, mechanisms, and rates of polyploid formation in flowering plants. Annu. Rev. Ecol. Syst. 29, 467–501. Rieseberg, L.H., and Willis, J.H. (2007). Plant speciation. Science 317, 910–914. Scott, R.J., Spielman, M., Bailey, J., and Dickinson, H.G. (1998). Parent-oforigin effects on seed development in Arabidopsis thaliana. Development 125, 3329–3341. Simon, P. (2003). Q-Gene: processing quantitative real-time RT-PCR data. Bioinformatics 19, 1439–1440. Smith, F.M., Garfield, A.S., and Ward, A. (2006). Regulation of growth and metabolism by imprinted genes. Cytogenet. Genome Res. 113, 279–291.
Kinoshita, T. (2007). Reproductive barrier and genomic imprinting in the endosperm of flowering plants. Genes Genet. Syst. 82, 177–186.
Tiwari, S., Spielman, M., Schulz, R., Oakey, R.J., Kelsey, G., Salazar, A., Zhang, K., Pennell, R., and Scott, R.J. (2010). Transcriptional profiles underlying parent-of-origin effects in seeds of Arabidopsis thaliana. BMC Plant Biol. 10, 72.
Ngo, Q.A., Moore, J.M., Baskar, R., Grossniklaus, U., and Sundaresan, V. (2007). Arabidopsis GLAUCE promotes fertilization-independent endosperm
Veitia, R.A. (2003). Nonlinear effects in macromolecular assembly and dosage sensitivity. J. Theor. Biol. 220, 19–25.
534 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Vrana, P.B., Fossella, J.A., Matteson, P., del Rio, T., O’Neill, M.J., and Tilghman, S.M. (2000). Genetic and epigenetic incompatibilities underlie hybrid dysgenesis in Peromyscus. Nat. Genet. 25, 120–124. Walia, H., Josefsson, C., Dilkes, B., Kirkbride, R., Harada, J., and Comai, L. (2009). Dosage-dependent deregulation of an AGAMOUS-LIKE gene cluster contributes to interspecific incompatibility. Curr. Biol. 19, 1128–1132. Wang, Y.J., and Brock, H.W. (2003). Polyhomeotic stably associates with molecular chaperones Hsc4 and Droj2 in Drosophila Kc1 cells. Dev. Biol. 262, 350–360. Weinhofer, I., Hehenberger, E., Roszak, P., Hennig, L., and Ko¨hler, C. (2010). H3K27me3 profiling of the endosperm implies exclusion of polycomb group protein targeting by DNA methylation. PLoS Genet. 6, e1001152.
Weijers, D., Geldner, N., Offringa, R., and Ju¨rgens, G. (2001). Seed development: Early paternal gene activity in Arabidopsis. Nature 414, 709–710. Wolff, P., Weinhofer, I., Seguin, J., Roszak, P., Beisel, C., Donoghue, M.T., Spillane, C., Nordborg, M., Rehmsmeier, M., and Ko¨hler, C. (2011). Highresolution analysis of parent-of-origin allelic expression in the Arabidopsis Endosperm. PLoS Genet. 7, e1002126. Woodell, S.R.J., and Valentine, D.H. (1961). Studies in British primulas. IX. Seed incompatibility in diploid-autotetraploid crosses. New Phytol. 60, 282–294. Zeh, D.W., and Zeh, J.A. (2000). Reproductive mode and speciation: the viviparity-driven conflict hypothesis. Bioessays 22, 938–946.
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 535
Article An Imprinted Gene Underlies Postzygotic Reproductive Isolation in Arabidopsis thaliana David Kradolfer,1,2 Philip Wolff,1,2 Hua Jiang,1 Alexey Siretskiy,1 and Claudia Ko¨hler1,* 1Department of Plant Biology and Forest Genetics, Uppsala BioCenter, Swedish University of Agricultural Sciences and Linnean Center of Plant Biology, 750 07 Uppsala, Sweden 2Department of Biology and Zu ¨ rich-Basel Plant Science Center, Swiss Federal Institute of Technology, ETH Centre, CH-8092 Zu¨rich, Switzerland *Correspondence: [email protected] http://dx.doi.org/10.1016/j.devcel.2013.08.006
SUMMARY
Postzygotic reproductive isolation in response to interploidy hybridizations is a well-known phenomenon in plants that forms a major path for sympatric speciation. A main determinant for the failure of interploidy hybridizations is the endosperm, a nutritious tissue supporting embryo growth, similar to the functional role of the placenta in mammals. Although it has been suggested that deregulated imprinted genes underpin dosage sensitivity of the endosperm, the molecular basis for this phenomenon remained unknown. In a genetic screen for suppressors of triploid seed abortion, we have identified the paternally expressed imprinted gene ADMETOS (ADM). Here, we present evidence that increased dosage of ADM causes triploid seed arrest. A large body of theoretical work predicted that deregulated imprinted genes establish the barrier to interploidy hybridization. Our study thus provides evidence strongly supporting this hypothesis and generates the molecular basis for our understanding of postzygotic hybridization barriers in plants.
INTRODUCTION How new species are formed remains a major question in evolutionary biology. Solving this question requires an understanding of the genetic and evolutionary basis of the traits mediating reproductive isolation. Postzygotic reproductive isolation in response to interploidy hybridization forms a major path for sympatric speciation in plants (Otto and Whitton, 2000; Ramsey and Schemske, 1998). The phenomenon has been termed ‘‘triploid block,’’ describing the difficulty of obtaining viable triploid seeds by diploid-tetraploid crosses (Marks, 1966). The endosperm is a key determinant for the success of interploidy hybridizations (Brink and Cooper, 1947; Woodell and Valentine, 1961), and abnormalities in the growth and structure of the endosperm are the source of the triploid block (Ramsey and Schemske, 1998). In most angiosperms, the endosperm is a triploid tissue derived after fertilization of the homodiploid central cell with a haploid sperm cell. The endosperm serves to support
and nurture the growing embryo, similar to the role of the placenta in mammals (Li and Berger, 2012). The endosperm of Arabidopsis thaliana follows the nuclear type of development where an initial syncytial phase of free nuclear divisions without cytokinesis is followed by cellularization (Costa et al., 2004). Endosperm cellularization is an important developmental transition required for viable seed formation (Hehenberger et al., 2012). Reciprocal hybridizations of plants that differ in ploidy have antagonistic effects on endosperm development; pollinations of maternal plants with pollen donors of higher ploidy will cause delay or failure of endosperm cellularization correlating with increased seed size or seed abortion. Conversely, the reciprocal cross will cause precocious endosperm cellularization and decreased seed size (Scott et al., 1998). In Arabidopsis, hybridizations of a diploid female plant with a tetraploid pollen donor (paternal excess hybridizations) will cause the formation of triploid progeny that will abort in an accession-dependent frequency. Although abortion of triploid seeds is highly penetrant in the Arabidopsis Columbia (Col) accession, the response to interploidy hybridization is rather weak in the Landsberg erecta (Ler) and C24 accessions (Scott et al., 1998; Dilkes et al., 2008). The Endosperm Balance Number (EBN) serves as a predictor for the success of interploidy hybridizations by assigning each species a defined effective ploidy that must be in a 2:1 maternal to paternal ratio in the endosperm for crosses to succeed (Johnston et al., 1980). Although widely used in plant breeding, the molecular mechanisms underpinning the EBN concept remained enigmatic. The sensitivity of the endosperm to changes in parental genome balance led to the hypothesis that imprinted genes are causal for the response to interploidy hybridizations (Haig and Westoby, 1989; Moore and Haig, 1991; Birchler, 1993; Gutierrez-Marcos et al., 2003; Zeh and Zeh, 2000; Kinoshita, 2007). Imprinted genes are predominantly expressed from either maternally or paternally inherited alleles as a result of epigenetic modifications established in the gametes (Ko¨hler et al., 2012). If imprinted genes encode for dosage-sensitive regulators, disturbed balance of these regulators in response to interploidy crosses is expected to cause deleterious effects (Birchler et al., 2001; Veitia, 2003; Birchler and Veitia, 2012). Although previous studies revealed deregulation of imprinted genes in response to interploidy hybridizations that were either connected with the activation of the silenced allele or hyperactivation of the active allele (Jullien and Berger, 2010; Tiwari et al., 2010; Wolff et al.,
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 525
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
2011), whether indeed there is a causal relationship between deregulated imprinted genes and seed abortion remained unknown. In this study, we have identified the paternally expressed imprinted gene (PEG) ADMETOS (ADM) to be causally responsible for abortion of triploid paternal excess Arabidopsis seeds. ADM acts as a dosage-sensitive regulator of triploid seed development and increased ADM expression in triploid seeds causes endosperm cellularization failure leading to seed abortion. Our study establishes ADM as a dosage-sensitive imprinted gene that erects a strong postzygotic reproductive barrier upon change of genome dosage. RESULTS Identification of a Suppressor of Triploid Seed Abortion To identify the genetic basis of interploidy hybridization barriers we searched for mutants that act as suppressors of triploid seed abortion. We made use of the jason (jas) mutant that frequently forms unreduced (2n) pollen and triploid seeds upon self-fertilization (Erilova et al., 2009). We generated a mutagenized population of diploid jas-3 plants (in the Col background) and screened for mutants with a reduced seed abortion rate. We identified one mutant that strongly suppressed triploid seed abortion and that we named admetos (adm-1) after one of Jason’s companions in his quest for the Golden Fleece. Although jas-3 formed about 30% aborted triploid seeds, only about 2% of seeds aborted in the jas-3 adm-1 double mutant (Figures 1A and 1C). We tested viability of jas-3 adm-1 seeds and found that 97% of seeds germinated compared with only 80% of jas-3 seeds (Figure 1B). Rescued triploid adm-1 seeds were enlarged compared to diploid wildtype seeds (Figure 1C). We tested ploidy of enlarged seeds and found 11 out of 12 tested seeds to be triploid (data not shown). We tested whether the ability of adm-1 to rescue triploid seeds is independent of the mutation giving rise to unreduced gametes. For that purpose, we introduced adm-1 into the omission of second division 1 (osd1) mutant background. In contrast to jas, which produces gametes by a first division restitution mechanism, the osd1 mutant forms unreduced gametes by a second division restitution mechanism (d’Erfurth et al., 2009). The osd1 mutation is highly penetrant and leads to almost 100% 2n pollen and triploid seeds when used as pollen donor (d’Erfurth et al., 2009). Similar to the rescue of triploid seeds in the jas-3 background, adm-1 rescued triploid seeds formed by unreduced pollen of the osd1 mutant (Figures 1D–1F). The majority of triploid adm-1 seeds formed by unreduced pollen of the osd1 mutant germinated and developed into seedlings (Figures 1E and 1G). In contrast to triploid seeds that fail to undergo endosperm cellularization (Scott et al., 1998), the endosperm of triploid adm-1 seeds cellularized, albeit delayed compared to diploid wild-type seeds (Figure 1H). Whereas wild-type seeds cellularize 5–7 days after pollination (DAP), triploid adm-1 seeds cellularized 8–10 DAP. Endosperm cellularization failure is likely responsible for embryo arrest (Hehenberger et al., 2012). In agreement with that view, restoration of endosperm cellularization in triploid adm-1 seeds allows progression of embryo development to the mature embryo stage (Figure 1H).
Expression of AGL MADS-Box Genes and Imprinted Genes Is Normalized in Triploid adm Seeds Previous studies revealed that failure of endosperm cellularization was associated with increased expression of AGAMOUS LIKE (AGL) MADS-box transcription factor genes (Erilova et al., 2009; Walia et al., 2009; Jullien and Berger, 2010; Hehenberger et al., 2012). Consistently, normalization of endosperm cellularization and embryo development by adm-1 was reflected by normalization of AGL gene expression (Figure 2A). Expression of AGL36, AGL62, and AGL90 was close to wild-type expression levels in triploid adm-1 seeds, whereas PHE1 and AGL40 had reduced expression levels in triploid adm-1 seeds compared to wild-type triploid seeds but still significantly increased expression compared to diploid wild-type seeds. Expression of AGL28 was not affected by the adm-1 mutation (Figure 2A). We also tested expression of MEA and FIS2 that are inversely deregulated in Col triploid seeds (Jullien and Berger, 2010). Whereas expression of MEA was reduced to wild-type levels in adm-1 triploid seeds, FIS2 expression was increased in adm-1 triploid seeds compared to diploid and triploid wild-type seeds (Figure 2B). Unbalanced parental contributions in the endosperm cause deregulated expression of imprinted genes (Jullien and Berger, 2010; Wolff et al., 2011). We therefore tested whether adm-1 could normalize expression of maternally and paternally expressed imprinted genes (MEGs and PEGs, respectively) that have increased expression levels in response to interploidy crosses (Wolff et al., 2011). Indeed, all tested PEGs had reduced expression levels in adm-1 triploid seeds compared to triploid seeds expressing a wild-type ADM allele (Figure 2C). Three out of six tested MEGs had increased expression levels in triploid seeds and their expression was decreased in adm-1 triploid seeds (Figure 2D). ADM Is a Paternally Expressed Imprinted Gene To examine whether adm-1 has a maternal or paternal effect on triploid seed rescue, we performed reciprocal crosses of adm-1 in the osd1 mutant background. These crosses revealed that adm-1 confers seed rescue when paternally but not when maternally inherited (Figures 3A and 3B). Loss of ADM affected only development of triploid but not diploid seeds. Seed development, final seed size, and expression of selected AGL genes were similar in diploid mutant seeds and wildtype seeds, suggesting a specific negative effect of increased ADM dosage in triploid seeds (Figures S1A–S1G available online). We identified the adm-1 mutation to cause an amino acid substitution in gene At4g11940, encoding for a protein belonging to the diverse family of molecular chaperones called J-domain proteins (Rajan and D’Silva, 2009; Kampinga and Craig, 2010). The mutation is located within the J-domain and changes an arginine at position 81 to a tryptophan, which likely impairs the function of the protein (Figures S2A and S2B). Consistently, genetic analysis revealed that adm-1 is likely a recessive mutant because diploid jas-3 pollen heterozygous for adm-1 did not produce viable triploid seeds (Figure S2C). Although all J-domain proteins share the name-giving conserved J-domain, they differ by the presence of different carboxyterminal domains. ADM belongs to the type IV class of J-domain proteins, which contain a J-domain lacking the
526 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 1. Triploid adm Seeds Are Viable (A and D) Percentage of aborted seeds of self-fertilized jas-3 and jas-3 adm-1 mutants (A) and Col wild-type plants pollinated with osd1 (D) and adm-1 mutants pollinated with osd1 adm-1 pollen (D). Numbers on top of bars correspond to number of analyzed seeds. (B and E) Percentage of germinated seeds of indicated genotypes. Numbers on top of bars correspond to number of analyzed seeds. (C and F) Pictures of mature seeds of indicated genotypes. Selected triploid seeds are marked with arrowheads. Scale bar, 1 mm. (G) Seedlings 6 days after germination. Scale bar, 1 cm. (H) Sections of seeds with the indicated genotype at 8 DAP (upper panels) and 10 DAP (lower panels). Ploidy was inferred based on seed morphology. Scale bar, 0.1 mm. See also Figures S1 and S2.
characteristic HPD motif required for stimulation of HSP70’s ATPase activity (Rajan and D’Silva, 2009; Kampinga and Craig, 2010). J-domain proteins are involved in multiple biological processes such as protein folding, protein translocation into
organelles, and regulation of gene expression by assisting chromatin regulators (Dai et al., 2005; Kampinga and Craig, 2010; Wang and Brock, 2003). Orthologous sequences to ADM were only identified in species of the Brassicaceae
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 527
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 2. Expression of AGL MADS-Box Genes and Imprinted Genes Is Decreased in adm Quantitative RT-PCR analysis of indicated genes in whole siliques derived from crosses of wild-type Col plants (white bars), wild-type Col plants pollinated with osd1 pollen (gray bars), and adm-1 plants pollinated with osd1 adm-1 pollen (black bars) at 6 DAP. (A) AGAMOUS LIKE (AGL) genes. (B) MEA and FIS2. (C) Paternally expressed imprinted genes. (D) Maternally expressed imprinted genes. Error bars indicate SEM.
closely related to Arabidopsis (A. lyrata, A. arenosa, and Capsella rubella; Figure S2A) but not in more distantly related species such as Brassica rapa nor in species outside the Brassicaceae. At4g11940 has previously been predicted to be a paternally expressed imprinted gene (Hsieh et al., 2011; Wolff et al., 2011), which is in agreement with our genetic data that reveal a functional role of ADM when paternally inherited. We introduced a construct containing the promoter and coding region of At4g1190 into the osd1 adm-1 background and tested whether it could complement the adm-1 phenotype. Indeed, the fraction of aborted seeds was increased to about 85% when we pollinated adm-1 plants with osd1
adm-1 pollen homozygous for the At4g11940 construct (Figure S2D). ADM was expressed during seed development, reaching highest expression levels at 3 DAP (Figure 3C). We generated a translational reporter of ADM with GFP and tested expression in pollen and developing seeds (Figures 3D–3K). The ADM::ADM-GFP transgene was functional as revealed by increased rates of seed abortion of triploid adm-1 seeds (Figure S2D). ADM was specifically localized in endosperm nuclei (Figures 3D–3K). Reciprocal crosses revealed that the ADM-GFP transgene was only expressed from the paternal allele (Figures 3H and 3I). We did not detect expression of ADM in pollen (Figures 3J and 3K) and neither in other tissues (data not shown), revealing that expression of the paternal ADM allele is specifically activated in the endosperm. ADM transcript levels were strongly increased in triploid seeds (Figure 3L). We tested whether increased ADM transcript levels are caused by breakdown of ADM imprinting and activation of the maternal ADM allele. However, ADM remained imprinted in triploid seeds (Figure 3M), revealing a specific effect of double paternal genome dosage on expression of the paternal ADM allele. ADM Is a Dosage-Sensitive Gene We addressed the question whether ADM is a dosage-sensitive gene, which would imply that triploid seed rescue correlates with decreasing ADM gene dosage. In agreement with this hypothesis, there was a decreased proportion of aborted triploid seeds when adm-1 was biparentally inherited compared to paternal inheritance (Figure 3B), suggesting a functional role of the maternal ADM alleles in triploid seeds. In agreement with that, there were residual levels of the maternal ADM alleles detectable in diploid and triploid seeds (Figure 3M;
528 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 3. ADM Is a Paternally Expressed Imprinted Gene and Remains Imprinted in Triploid Seeds (A) Pictures of mature seeds of adm-1 plants pollinated with osd1 pollen and wild-type Col plants pollinated with osd1 adm-1 pollen. (B) Percentages of aborted and germinated seeds of wild-type Col plants pollinated with osd1 pollen (white bars), adm-1 plants pollinated with osd1 pollen (light gray bars), wild-type Col plants pollinated with osd1 adm-1 pollen (dark gray bars), and adm-1 plants pollinated with osd1 adm-1 pollen (black bars). Numbers on top of bars correspond to number of analyzed seeds. (C) Expression of ADM in whole siliques from 1–6 DAP. Error bars indicate SEM. (D–G) ADM::ADM-GFP expression in self-fertilized seeds at 1, 2, 3, and 4 DAP. (H and I) Reciprocal crosses with ADM::ADM-GFP transmitted through pollen (H) and female gametes (I) at 4 DAP. (J and K) Only background fluorescence is detectable in ADM::ADM-GFP pollen (J) and wild-type pollen (K). (L) Expression of ADM in whole siliques derived from crosses of wild-type Col plants and wild-type plants pollinated with osd1 pollen at 6 DAP. Error bars indicate SEM. (M) ADM is predominantly paternally expressed in diploid and triploid seeds. Siliques of crosses of Ler plants pollinated with Col wild-type or jas-3 pollen were harvested at the indicated time points and allele-specific expression was tested by restriction-based allele-specific PCR analysis. Siliques of Col and Ler plants pollinated with Col and Ler pollen were used as controls. Scale bars, 1 mm (A) and 50 mm (D–K). See also Figure S3.
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 529
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 4. ADM Is a Dosage-Sensitive Gene (A) Distribution of seed sizes from self-fertilized wild-type Col plants (blue line), wild-type Col plants pollinated with osd1 pollen (red line), adm-1 plants pollinated with osd1 pollen (green line), wild-type Col plants pollinated with osd1 adm-1 pollen (purple line), and adm-1 plants pollinated with osd1 adm-1 pollen (black line). A minimum of 100 seeds was analyzed for each cross. (B) Frequencies of aborted and germinated triploid seeds derived from wild-type Col plants pollinated with osd1 pollen (light gray bars), wildtype Col plants pollinated with osd1 adm-2 pollen (dark gray bars), and adm-2 plants pollinated with osd1 adm-2 pollen (black bars). Numbers above the bars indicate the number of analyzed seeds. (C) Pictures of seeds of the indicated crosses. Scale bar, 1 mm. (D) Expression of AGL genes in whole siliques of indicated genotypes at 6 DAP. Error bars indicate SEM. See also Figure S4.
Figure S3). A functional role of maternal ADM expression, however, became unmasked only upon loss of paternal ADM alleles. Triploid seeds that were homozygous for adm-1 were decreased in size compared to triploid seeds inheriting a wild-type maternal ADM allele (genotype adm-1/ /+), revealing a dosage-sensitive role of ADM in regulating seed size (Figure 4A). We identified a second adm allele (adm-2; Figures S4A–S4C) in a collection of T-DNA mutants that similarly to adm-1 strongly suppressed triploid seed abortion (Figures 4B and 4C). In agreement with ADM being a dosage-sensitive gene, also adm-2 had a stronger effect on triploid seed viability when the seeds were homozygous for the adm-2 mutation compared to triploid seeds having a functional ADM maternal
allele (Figures 4B and 4C). We tested whether ADM dosage-dependent abortion of triploid seeds was reflected on transcript levels of AGL genes. Indeed, transcript levels of AGL genes in triploid adm-2 seeds were partially normalized (Figure 4D) with the clear trend that AGL expression was more strongly reduced in homozygous adm-2 triploid seeds than in triploid seeds that inherited an ADM maternal allele. The repressive effect of adm-2 on AGL gene expression was detectable at 4 DAP (Figure S4D), clearly preceding the onset of endosperm cellularization at 6 DAP (Kradolfer et al., 2013). Therefore, reduced AGL gene expression by adm in triploid seeds is unlikely a secondary consequence of normalized endosperm cellularization, but rather enables endosperm cellularization. Taken together, these data reveal that ADM is a dosage-sensitive gene and show that the maternal ADM allele gains a functional role in triploid seeds. We addressed the question whether ADM overexpression would be sufficient to confer diploid seed abortion. We generated transgenic lines overexpressing ADM under the control of the endosperm-specific PHE1 promoter (Weinhofer et al., 2010) and the RPS5A promoter, which is expressed in embryo and endosperm (Weijers et al., 2001). We generated 10 independent transgenic lines for each construct and identified several lines with medium to strong overexpression in the endosperm (Figure S4E). However, in none of these lines did we detect triploid seed-like abortion, revealing that increased ADM expression is specifically detrimental in triploid seeds.
530 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
in triploid Ler seeds compared to triploid Col seeds (Figure 5A), correlating with a reduced abortion rate of triploid Ler seeds. Introducing a Ler genomic locus of ADM (including 1.5 kb of upstream and 0.3 kb of downstream sequence) into the adm-1 mutant background (Col accession) complemented the mutant similarly well as an ADM genomic fragment derived from the Col accession (Figure 5B), revealing that ADM sequence variation between Col and Ler accessions (Figure S5) is unlikely to be responsible for accession-dependent differences in triploid seed abortion. In support of this conclusion, expression of a Col ADM genomic fragment in the jas-1 mutant (Ler accession) did not increase the frequency of triploid seed abortion (Figure 5C). Together, these data strongly suggest that the different triploid responses of Col and Ler are unlikely consequences of ADM sequence variation in cis but rather are caused by accession-specific factors acting in trans on ADM expression.
Figure 5. Accession-Dependent Triploid Seed Abortion Correlates with ADM Expression Levels (A) Expression of ADM in whole siliques derived from pollinations of wild-type Col or Ler plants with wild-type Col or Ler pollen (white and dark gray bars, respectively), Col plants pollinated with jas-3 pollen (Col background; light gray bars), and Ler plants pollinated with jas-1 pollen (Ler background; black bars) at 4 and 6 DAP. Error bars indicate SEM. (B) adm-1 plants were pollinated with pollen of osd1, osd1 adm-1, or osd1 adm-1 mutants hemizygous for a Col or Ler ADM::ADM transgene. Percentage of collapsed seeds is shown. Numbers above the bars indicate the number of analyzed seeds. Independent transgenic lines are numbered (#). (C) A Col ADM::ADM transgene was expressed in the jas-1 (Ler) background. Percentage of collapsed seeds for self-fertilized plants is shown. Numbers above the bars indicate the number of analyzed seeds. Independent transgenic lines are numbered (#). See also Figure S5.
Accession-Dependent Seed Abortion Correlates with ADM Expression Levels Triploid seed abortion is accession dependent. Although triploid seeds abort very frequently in the Col accession, the majority of triploid seeds are viable in the Ler background (Dilkes et al., 2008). We addressed the question whether accession-dependent differences of Col and Ler are caused by different expression or function of ADM. Therefore, we tested whether triploid Col and Ler seeds differ in ADM expression levels. Indeed, we found that ADM transcript levels remained at a much lower level
adm Partially Rescues Seeds Lacking FIS-PRC2 Loss of function of the FERTILIZATION INDEPENDENT SEED (FIS) Polycomb Repressive Complex 2 (PRC2) causes a similar phenotype as paternal excess triploid seeds, correlating with largely overlapping sets of deregulated genes (Erilova et al., 2009; Tiwari et al., 2010). Consistently, ADM is a target gene of the FIS-PRC2 in the endosperm (Figure S6), and ADM expression was increased in mutants of the FIS-PRC2 subunit MEA (Figure 6A). We tested whether similar to its ability to rescue triploid seeds adm-1 could rescue mea mutant seeds. We generated double mutants that were heterozygous for mea and homozygous for adm-1 and pollinated them with adm-1 mutant pollen. Whereas a mea/+ ADM/+ mutant formed about 50% aborting seeds (n = 216), seed abortion in mea/+ adm-1/ double mutants was suppressed to 36% (n = 357; X2 = 29.14 > X0.05[1]2 = 3.84). Among seeds produced by a mea/+ adm-1/ mutant, we found 10% partially collapsed seeds that contained a developed embryo (Figures 6B and 6C). We tested viability of aborted and partially collapsed seeds by germinating them on plates and genotyping the seedlings for the presence of the mea mutation. We found that 15% of mea/+ adm-1/ seeds germinated, compared to only 1%–3% of mea/+ and mea/+ adm-1/+ seeds (Figure 6D). These data reveal that loss of ADM function allows the requirement of FIS-PRC2 function to be bypassed. DISCUSSION Here, we report the identification of the imprinted ADM gene, which determines seed viability in paternal excess interploidy hybridizations. Whereas triploid seeds abort at a frequency close to 100% in the Col accession, mutation of ADM leads to nearly complete seed rescue. Triploid adm seeds have reduced expression levels of AGL genes, including AGL62, a central regulator of endosperm cellularization. Consistently, cellularization is restored in triploid adm seeds. The parental conflict theory predicts that genomic imprinting evolved as a consequence of intragenomic conflict over nutrient allocation from the mother to the offspring (Haig and Westoby, 1989). Based on this prediction, PEGs should have a predominant role in supporting growth of tissues that sustain embryo growth (placenta and endosperm) (Haig and Westoby, 1989;
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 531
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Figure 6. adm Partially Suppresses Abortion of mea Mutant Seeds (A) Quantitative RT-PCR analysis of ADM in whole siliques derived from pollinations of wild-type Col plants or mea/+ mutants with wild-type Col pollen at 6 DAP. Error bars indicate SEM. (B) Percentages of aborted, partially collapsed, and normal seeds derived from pollinations of mea/+ with wild-type Col pollen, mea/+ adm-1 with wild-type Col pollen, mea/+ with adm-1 pollen, and mea/+ adm-1 with adm-1 pollen (left). Numbers above the bars indicate the number of analyzed seeds. Examples of seeds of the three categories (right panels). Scale bar, 0.5 mm. (C) Clearings of seeds derived from pollinations of mea/+ adm-1 with adm-1 pollen at 12 DAP. Wild-type seed (left), aborting mea seed (middle), and rescued mea seed (right) are segregated. Scale bar, 0.1 mm. (D) Transmission frequency of the mea mutant allele through female gametes was determined by genotyping germinated seedlings for the presence of the mea mutation. See also Figure S6.
Smith et al., 2006). Our results provide substantial support to this hypothesis by revealing that the PEG ADM has a role in promoting endosperm growth by regulating endosperm cellularization in triploid seeds. The effect on endosperm cellularization is likely a direct or indirect consequence of ADM, regulating expression of AGL genes in triploid seeds. Previous studies revealed that reduced paternal genome contribution can bypass the requirement of the FIS-PRC2 for viable seed formation (Nowack et al., 2007; Kradolfer et al., 2013). Here, we show that reduced dosage of ADM can substantially contribute to fis mutant rescue, thus providing strong support for the view that the FIS-PRC2 regulates viable seed formation by regulating expression of PEGs. Whereas increased expression of ADM is causally responsible for triploid seed abortion, increased expression of ADM did not cause major changes in diploid seed development. These data reveal that changes in ADM dosage have different consequences in diploid and triploid seeds. Moreover, impaired ADM function did not cause noticeable aberrations in diploid seed development, suggesting that there is either residual ADM function in adm-1 and adm-2 mutants, or, alternatively, that ADM is not essential for seed development. The last scenario bears
striking similarities to the lethal effect mediated by Lethal hybrid rescue (Lhr) in F1 hybrid sons of Drosophila melanogaster and D. simulans (Brideau et al., 2006). A loss-of-function mutation at the D. simulans Lhr locus can reverse F1 lethality; however, neither does increased expression of the D. simulans Lhr cause lethality in D. melanogaster nor does loss of Lhr function cause a negative effect (Chatterjee et al., 2007; Brideau and Barbash, 2011). Therefore, the Lhr protein acquires a negative ectopic interaction in the hybrid background, similar to the negative effect of ADM in triploid seeds. Speciation is characterized by the evolution of barriers to reproductive exchange between two groups of organisms. The genes building these hybridization barriers are speciation genes. Our study identifies ADM as an imprinted plant speciation gene that underlies the hybridization barrier between diploid and tetraploid Arabidopsis plants. Although only a few natural Arabidopsis thaliana tetraploid accessions have been identified (Bomblies et al., 2007), unreduced male gametes are frequently formed after cold stress (De Storme et al., 2012). It is generally assumed that the formation of stable tetraploids occurs via the formation of unreduced gametes and unstable triploid intermediates (Rieseberg and Willis, 2007). We
532 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
therefore propose that suppression of triploid seed formation by ADM contributes to the low frequency of natural Arabidopsis thaliana tetraploid populations. However, because the triploid block mediated by ADM is active only in response to increased paternal genome contribution, triploid seed formation can occur when unreduced gametes are maternally formed, explaining the existence of tetraploid populations of Arabidopsis thaliana and relatives. A large body of theoretical work led to the prediction that imprinted genes erect hybridization barriers (Haig and Westoby, 1989; Moore and Haig, 1991; Birchler, 1993; Zeh and Zeh, 2000; Gutierrez-Marcos et al., 2003; Kinoshita, 2007). This theory builds on the hypothesis that imprinted genes are dosage sensitive and imprinted expression is disrupted in response to hybridization. Disruption of imprinted gene expression has been shown to occur at some selected loci in response to interploidy and interspecies hybridization in Arabidopsis (Josefsson et al., 2006; Jullien and Berger, 2010) as well as in the North American rodent species Peromyscus (Vrana et al., 2000). Our study shows that interploidy hybridizations result in hyperactivation of the active ADM allele while not affecting the imprinted status of ADM. Until now experimental evidence that unbalanced expression of an imprinted gene can form a hybridization barrier was missing. Thus, the identification of ADM as a dosage-sensitive imprinted gene represents a major advance because it is an imprinted plant gene that erects a strong postzygotic reproductive barrier upon change of genome dosage. One implication of our data is that alterations in epigenetic gene regulation contribute to both the establishment and maintenance of reproductive isolation barriers in plants. EXPERIMENTAL PROCEDURES Plant Material and Growth Conditions Plants were grown in a growth chamber at 60% humidity and daily cycles of 16 hr light at 21 C and 8 hr darkness at 18 C. Arabidopsis thaliana mutants jas-3 (Erilova et al., 2009), mea-8 (Ngo et al., 2007), and adm-2 (SAIL_1252_D03) are in the Col accession. The osd1-1 mutant (d’Erfurth et al., 2009) was kindly provided by Raphael Mercier. The mutant was originally identified in the Nossen background and subsequently introgressed into Col by repeated backcrossing over five generations. The jas-1 mutant is in the Ler background (Erilova et al., 2009). For crosses, designated female partners were emasculated, and the pistils hand-pollinated 2 days after emasculation. RNA Extraction and qPCR Analysis Three siliques were harvested for each cross and frozen together with glass beads (1.25–1.55 mm; Carl Roth) in liquid nitrogen. The samples were ground in a Silamat S5 (IvoclarVivadent) two times for 6 s. RNA was extracted using the RNeasy Plant Mini Kit (QIAGEN) according to the manufacturer’s instructions. Residual DNA was removed using the QIAGEN RNase-free DNase Set, and cDNA was synthesized using the Fermentas First strand cDNA synthesis kit according to the manufacturer’s instructions. Quantitative real-time PCR was performed using an iQ5 Real-Time PCR Detection System (BioRad) and Maxima SYBR green qPCR master mix (Fermentas) according to the manufacturer’s instructions. Quantitative realtime PCR was performed with three replicates using primers as indicated in the Supplemental Experimental Procedures, and results were analyzed as described (Simon, 2003) using PP2A as a reference gene. EMS Mutagenesis and Mapping Diploid jas-3 seeds were selected based on seed size and mutagenized for 15 hr in 0.3% ethyl methanesulfonate (EMS). The seeds were germinated
on soil, and one side branch of each plant was harvested. We screened 900 M2 families consisting each of eight plants derived from one M1 plant and selected mutants with reduced seed abortion. For genetic mapping of the adm mutation, we established an F2 mapping population by crossing jas-3 adm with jas-3. From 896 analyzed plants, 200 had a clear adm phenotype and were selected for mapping. Leaf samples of all plants were pooled and DNA was extracted using the Nucleon PhytoPure Kit (Amersham Biosciences). Sequencing was performed on one lane of an Illumina Hiseq 2000 device using paired end mode at 70 fold coverage. Read alignment was done using the Bowtie algorithm (Langmead et al., 2009), and Samtools was employed for SNP calling (Li et al., 2009). Mapping using SNP data files was performed as described by (Austin et al., 2011). Microscopy For tissue sections, seeds were fixed and embedded with Technovit 7100 (Heraeus Kulzer) as described (Hehenberger et al., 2012). Five-micrometer sections were prepared with an HM 355 S microtome (Microm) using glass knives. Sections were stained for 1 min with 0.1% toluidine blue and washed three times with distilled water. For clearings, siliques were slit along the lamella and fixed overnight at 4 C in ethanol/acetic acid in a ratio of 9:1 and washed with 70% ethanol. Seeds were cleared for 1 day at 4 C in chloral hydrate solution (glycerol/chloral hydrate/water in a ratio of 1:8:3). Seeds and pollen of the GFP lines were mounted in water and analyzed with epifluorescence optics. Microscopy was performed using a Leica DMI 4000B microscope with DIC optics. Images were captured using a Leica DFC360 FX camera (Leica) and processed using Adobe Photoshop CS5. Pictures of mature seeds were taken using a Leica Z16apoA microscope and a Leica DFC425C camera and processed using Adobe Photoshop CS5. Flow Cytometry Ploidy levels were measured as described (Kradolfer et al., 2013). Seed Size Analysis Seeds were arranged on glass slides and pictures were taken using a Leica Z16apoA microscope with transmitted light. Seed size was measured as described by (Herridge et al., 2011). Germination and Transmission Analysis Seeds were surface sterilized using chlorine gas. Five milliliters of hydrochloric acid was added to 30 ml of sodium hypochlorite under a fume hood. This mixture was incubated in a container together with the seeds for 1 hr. Seeds were plated on MS media containing 1% sugar. After stratification at 4 C in the dark for 2 days, seedlings were grown in a growth room under a longday photoperiod (16 hr light and 8 hr darkness) at 23 C. Germination frequency was determined after 7 days. For transmission analysis, seedlings were harvested after 12 days for genotyping using primers specified in the Supplemental Experimental Procedures. Generation of Plasmids and Transgenic Lines Details on the generation of plasmids and transgenic lines are provided in the Supplemental Experimental Procedures. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures and six figures and can be found with this article online at http://dx.doi.org/ 10.1016/j.devcel.2013.08.006. ACKNOWLEDGMENTS We thank Sabrina Huber and Cecilia Wa¨rdig for excellent technical support, Sofia Berlin Kolm and Juliette de Meaux for helpful discussions, and Lars Hennig and members of the Ko¨hler laboratory for critical comments on this manuscript. This research was supported by a fellowship of the Zu¨rich-Basel Plant Science Center (to D.K.) and a European Research Council Starting Independent Researcher grant (to C.K.).
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 533
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Received: April 3, 2013 Revised: June 18, 2013 Accepted: August 10, 2013 Published: September 5, 2013 REFERENCES Austin, R.S., Vidaurre, D., Stamatiou, G., Breit, R., Provart, N.J., Bonetta, D., Zhang, J., Fung, P., Gong, Y., Wang, P.W., et al. (2011). Next-generation mapping of Arabidopsis genes. Plant J. 67, 715–725. Birchler, J.A. (1993). Dosage analysis of maize endosperm development. Annu. Rev. Genet. 27, 181–204. Birchler, J.A., Bhadra, U., Bhadra, M.P., and Auger, D.L. (2001). Dosagedependent gene regulation in multicellular eukaryotes: implications for dosage compensation, aneuploid syndromes, and quantitative traits. Dev. Biol. 234, 275–288. Birchler, J.A., and Veitia, R.A. (2012). Gene balance hypothesis: connecting issues of dosage sensitivity across biological disciplines. Proc. Natl. Acad. Sci. USA 109, 14746–14753. Bomblies, K., Lempe, J., Epple, P., Warthmann, N., Lanz, C., Dangl, J.L., and Weigel, D. (2007). Autoimmune response as a mechanism for a DobzhanskyMuller-type incompatibility syndrome in plants. PLoS Biol. 5, e236. Brideau, N.J., and Barbash, D.A. (2011). Functional conservation of the Drosophila hybrid incompatibility gene Lhr. BMC Evol. Biol. 11, 57. Brideau, N.J., Flores, H.A., Wang, J., Maheshwari, S., Wang, X., and Barbash, D.A. (2006). Two Dobzhansky-Muller genes interact to cause hybrid lethality in Drosophila. Science 314, 1292–1295. Brink, R., and Cooper, D. (1947). The endosperm in seed development. Bot. Rev. 132, 479–541. Chatterjee, R.N., Chatterjee, P., Pal, A., and Pal-Bhadra, M. (2007). Drosophila simulans Lethal hybrid rescue mutation (Lhr) rescues inviable hybrids by restoring X chromosomal dosage compensation and causes fluctuating asymmetry of development. J. Genet. 86, 203–215. Costa, L.M., Gutie`rrez-Marcos, J.F., and Dickinson, H.G. (2004). More than a yolk: the short life and complex times of the plant endosperm. Trends Plant Sci. 9, 507–514. Dai, Y.S., Xu, J., and Molkentin, J.D. (2005). The DnaJ-related factor Mrj interacts with nuclear factor of activated T cells c3 and mediates transcriptional repression through class II histone deacetylase recruitment. Mol. Cell. Biol. 25, 9936–9948. De Storme, N., Copenhaver, G.P., and Geelen, D. (2012). Production of diploid male gametes in Arabidopsis by cold-induced destabilization of postmeiotic radial microtubule arrays. Plant Physiol. 160, 1808–1826. d’Erfurth, I., Jolivet, S., Froger, N., Catrice, O., Novatchkova, M., and Mercier, R. (2009). Turning meiosis into mitosis. PLoS Biol. 7, e1000124. Dilkes, B.P., Spielman, M., Weizbauer, R., Watson, B., Burkart-Waco, D., Scott, R.J., and Comai, L. (2008). The maternally expressed WRKY transcription factor TTG2 controls lethality in interploidy crosses of Arabidopsis. PLoS Biol. 6, 2707–2720. Erilova, A., Brownfield, L., Exner, V., Rosa, M., Twell, D., Mittelsten Scheid, O., Hennig, L., and Ko¨hler, C. (2009). Imprinting of the polycomb group gene MEDEA serves as a ploidy sensor in Arabidopsis. PLoS Genet. 5, e1000663. Gutierrez-Marcos, J.F., Pennington, P.D., Costa, L.M., and Dickinson, H.G. (2003). Imprinting in the endosperm: a possible role in preventing wide hybridization. Philos. Trans. R. Soc. Lond. B Biol. Sci. 358, 1105–1111. Haig, D., and Westoby, M. (1989). Parent-specific gene expression and the triploid endosperm. Am. Nat. 134, 147–155. Hehenberger, E., Kradolfer, D., and Ko¨hler, C. (2012). Endosperm cellularization defines an important developmental transition for embryo development. Development 139, 2031–2039.
development and expression of paternally inherited alleles. Development 134, 4107–4117. Herridge, R.P., Day, R.C., Baldwin, S., and Macknight, R.C. (2011). Rapid analysis of seed size in Arabidopsis for mutant and QTL discovery. Plant Methods 7, 3. Hsieh, T.F., Shin, J., Uzawa, R., Silva, P., Cohen, S., Bauer, M.J., Hashimoto, M., Kirkbride, R.C., Harada, J.J., Zilberman, D., and Fischer, R.L. (2011). Regulation of imprinted gene expression in Arabidopsis endosperm. Proc. Natl. Acad. Sci. USA 108, 1755–1762. Johnston, S., Nijs, T., Peloquin, S., and Hanneman, R., Jr. (1980). The significance of genic balance to endosperm development in interspecific crosses. Theor. Appl. Genet. 57, 5–9. Josefsson, C., Dilkes, B., and Comai, L. (2006). Parent-dependent loss of gene silencing during interspecies hybridization. Curr. Biol. 16, 1322–1328. Jullien, P.E., and Berger, F. (2010). Parental genome dosage imbalance deregulates imprinting in Arabidopsis. PLoS Genet. 6, e1000885. Kampinga, H.H., and Craig, E.A. (2010). The HSP70 chaperone machinery: J proteins as drivers of functional specificity. Nat. Rev. Mol. Cell Biol. 11, 579–592. Ko¨hler, C., Wolff, P., and Spillane, C. (2012). Epigenetic mechanisms underlying genomic imprinting in plants. Annu. Rev. Plant Biol. 63, 331–352. Kradolfer, D., Hennig, L., and Ko¨hler, C. (2013). Increased maternal genome dosage bypasses the requirement of the FIS polycomb repressive complex 2 in Arabidopsis seed development. PLoS Genet. 9, e1003163. Langmead, B., Trapnell, C., Pop, M., and Salzberg, S.L. (2009). Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25. Li, J., and Berger, F. (2012). Endosperm: food for humankind and fodder for scientific discoveries. New Phytol. 195, 290–305. Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., and Durbin, R.; 1000 Genome Project Data Processing Subgroup. (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079. Marks, G. (1966). The origin and significance of intraspecific polyploidy: experimental evidence from Solanum chacoense. Evolution 20, 552–557. Moore, T., and Haig, D. (1991). Genomic imprinting in mammalian development: a parental tug-of-war. Trends Genet. 7, 45–49. Nowack, M.K., Shirzadi, R., Dissmeyer, N., Dolf, A., Endl, E., Grini, P.E., and Schnittger, A. (2007). Bypassing genomic imprinting allows seed development. Nature 447, 312–315. Otto, S.P., and Whitton, J. (2000). Polyploid incidence and evolution. Annu. Rev. Genet. 34, 401–437. Rajan, V.B., and D’Silva, P. (2009). Arabidopsis thaliana J-class heat shock proteins: cellular stress sensors. Funct. Integr. Genomics 9, 433–446. Ramsey, J., and Schemske, D. (1998). Pathways, mechanisms, and rates of polyploid formation in flowering plants. Annu. Rev. Ecol. Syst. 29, 467–501. Rieseberg, L.H., and Willis, J.H. (2007). Plant speciation. Science 317, 910–914. Scott, R.J., Spielman, M., Bailey, J., and Dickinson, H.G. (1998). Parent-oforigin effects on seed development in Arabidopsis thaliana. Development 125, 3329–3341. Simon, P. (2003). Q-Gene: processing quantitative real-time RT-PCR data. Bioinformatics 19, 1439–1440. Smith, F.M., Garfield, A.S., and Ward, A. (2006). Regulation of growth and metabolism by imprinted genes. Cytogenet. Genome Res. 113, 279–291.
Kinoshita, T. (2007). Reproductive barrier and genomic imprinting in the endosperm of flowering plants. Genes Genet. Syst. 82, 177–186.
Tiwari, S., Spielman, M., Schulz, R., Oakey, R.J., Kelsey, G., Salazar, A., Zhang, K., Pennell, R., and Scott, R.J. (2010). Transcriptional profiles underlying parent-of-origin effects in seeds of Arabidopsis thaliana. BMC Plant Biol. 10, 72.
Ngo, Q.A., Moore, J.M., Baskar, R., Grossniklaus, U., and Sundaresan, V. (2007). Arabidopsis GLAUCE promotes fertilization-independent endosperm
Veitia, R.A. (2003). Nonlinear effects in macromolecular assembly and dosage sensitivity. J. Theor. Biol. 220, 19–25.
534 Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc.
Developmental Cell Postzygotic Reproductive Isolation in Arabidopsis
Vrana, P.B., Fossella, J.A., Matteson, P., del Rio, T., O’Neill, M.J., and Tilghman, S.M. (2000). Genetic and epigenetic incompatibilities underlie hybrid dysgenesis in Peromyscus. Nat. Genet. 25, 120–124. Walia, H., Josefsson, C., Dilkes, B., Kirkbride, R., Harada, J., and Comai, L. (2009). Dosage-dependent deregulation of an AGAMOUS-LIKE gene cluster contributes to interspecific incompatibility. Curr. Biol. 19, 1128–1132. Wang, Y.J., and Brock, H.W. (2003). Polyhomeotic stably associates with molecular chaperones Hsc4 and Droj2 in Drosophila Kc1 cells. Dev. Biol. 262, 350–360. Weinhofer, I., Hehenberger, E., Roszak, P., Hennig, L., and Ko¨hler, C. (2010). H3K27me3 profiling of the endosperm implies exclusion of polycomb group protein targeting by DNA methylation. PLoS Genet. 6, e1001152.
Weijers, D., Geldner, N., Offringa, R., and Ju¨rgens, G. (2001). Seed development: Early paternal gene activity in Arabidopsis. Nature 414, 709–710. Wolff, P., Weinhofer, I., Seguin, J., Roszak, P., Beisel, C., Donoghue, M.T., Spillane, C., Nordborg, M., Rehmsmeier, M., and Ko¨hler, C. (2011). Highresolution analysis of parent-of-origin allelic expression in the Arabidopsis Endosperm. PLoS Genet. 7, e1002126. Woodell, S.R.J., and Valentine, D.H. (1961). Studies in British primulas. IX. Seed incompatibility in diploid-autotetraploid crosses. New Phytol. 60, 282–294. Zeh, D.W., and Zeh, J.A. (2000). Reproductive mode and speciation: the viviparity-driven conflict hypothesis. Bioessays 22, 938–946.
Developmental Cell 26, 525–535, September 16, 2013 ª2013 Elsevier Inc. 535